linux-kernel - Re: [PATCH v2 08/12] powerpc: Don't use 'struct ppc

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <871r93vqcb.fsf@mpe.ellerman.id.au>
Date:   Tue, 15 Jun 2021 12:01:08 +1000
From:   Michael Ellerman <mpe@...erman.id.au>
To:     Christophe Leroy <christophe.leroy@...roup.eu>,
        Benjamin Herrenschmidt <benh@...nel.crashing.org>,
        Paul Mackerras <paulus@...ba.org>,
        naveen.n.rao@...ux.vnet.ibm.com, jniethe5@...il.com
Cc:     linux-kernel@...r.kernel.org, linuxppc-dev@...ts.ozlabs.org
Subject: Re: [PATCH v2 08/12] powerpc: Don't use 'struct ppc_inst' to
 reference instruction location

Christophe Leroy <christophe.leroy@...roup.eu> writes:
> diff --git a/arch/powerpc/include/asm/inst.h b/arch/powerpc/include/asm/inst.h
> index 5a0740ebf132..32d318c3b180 100644
> --- a/arch/powerpc/include/asm/inst.h
> +++ b/arch/powerpc/include/asm/inst.h
> @@ -139,7 +139,7 @@ static inline int ppc_inst_len(struct ppc_inst x)
>   * Return the address of the next instruction, if the instruction @value was
>   * located at @location.
>   */
> -static inline struct ppc_inst *ppc_inst_next(void *location, struct ppc_inst *value)
> +static inline unsigned int *ppc_inst_next(unsigned int *location, unsigned int *value)
>  {
>  	struct ppc_inst tmp;
>  

It's not visible in the diff, but the rest of the function is:

	tmp = ppc_inst_read(value);

	return location + ppc_inst_len(tmp);
}

And so changing the type of location from void * to int * changes the
result of that addition, ie. previously it was in units of bytes, now
it's units of 4 bytes.

To fix it I've kept location as unsigned int *, and added a cast where
we do the addition. That way users of the function just see unsigned int *,
the cast to void * is an implementation detail.

We only have a handful of uses of ppc_inst_len(), so maybe that should
change name and return a result in units of int *. But that can be a
separate change.

> diff --git a/arch/powerpc/platforms/86xx/mpc86xx_smp.c b/arch/powerpc/platforms/86xx/mpc86xx_smp.c
> index 87f524e4b09c..302f2a1e0361 100644
> --- a/arch/powerpc/platforms/86xx/mpc86xx_smp.c
> +++ b/arch/powerpc/platforms/86xx/mpc86xx_smp.c
> @@ -83,7 +83,7 @@ smp_86xx_kick_cpu(int nr)
>  		mdelay(1);
>  
>  	/* Restore the exception vector */
> -	patch_instruction((struct ppc_inst *)vector, ppc_inst(save_vector));
> +	patch_instruction(vector, ppc_inst(save_vector));
>  
>  	local_irq_restore(flags);
>  

There was another usage in here:

 	/* Setup fake reset vector to call __secondary_start_mpc86xx. */
 	target = (unsigned long) __secondary_start_mpc86xx;
-	patch_branch((struct ppc_inst *)vector, target, BRANCH_SET_LINK);
+	patch_branch(vector, target, BRANCH_SET_LINK);
 
 	/* Kick that CPU */
 	smp_86xx_release_core(nr);

I fixed it up.

cheers